Hi Michelle,
the text file stores the power at different frequencies - simply
encoded as power at different cycles. If your experiment is 100
volumes long then the highest frequency that can be represented is
the one that corresponds to 50 full cycles, independent of the TR.
The lowest frequency corresponds to a single cycle, i.e. all the
entries in the text file corresponds to a frequency related to the
number of cycles, the first value is the power for 1 full cycle and
the 50th value is the power for 50 full cycles. For a blocked design
this is the easiest way of reading the fft plots, i.e. if during the
course of your experiment you had 6 experimental blocks then the 6th
value in the text file should be large.
You can easily transform #cycles to frequencies, e.g. with 6 full
cycles in 100 vols and a TR of 3sec the corresponding frequency in Hz is
100 * 3sec / 6 = 50 sec per cycle, e.g. a frequency of 1 / 50 Hz
hope this helps
christian
On 6 Jun 2007, at 04:33, Michelle Voss wrote:
> the attachment..
>
> On 6/5/07, Michelle Voss <[log in to unmask]> wrote: Hi Everyone,
>
> We have a REALLY messy data set where ICA pulls out up to 400
> components.
> There is a large range of frequencies for these components, and the
> numbers
> overlap along the x-axis of the frequency graph. I've attached an
> example
> to show what i mean.
>
> I can't figure out how to rescale the power/frequency plot to be
> able to
> tell where the frequencies drop off. This will be a major
> indicator of
> whether the component is a possible keeper or noise..but i can't
> tell what
> the numbers are???
>
> The text file lists only the arbitrary power values, but i don't
> know how to
> get the frequencies on the x axis? as we could easily re-plot
> these w/out
> re-running ICA.
>
> am i missing something basic? can anyone tell me how to do this?
>
> thanks!
> Michelle
>
> <f1.png>
____
Christian F. Beckmann
University Research Lecturer
Oxford University Centre for Functional MRI of the Brain (FMRIB)
John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK.
[log in to unmask] http://www.fmrib.ox.ac.uk/~beckmann
tel: +44 1865 222551 fax: +44 1865 222717
|